This is a reland of b772a8f14e0aa806713f9042b2a8e186bcfc103d
Original patch: https://crrev.com/c/2341163
Revert: https://crrev.com/c/2348532

Reason for revert: New WPT test revealed a memory issue in
//third_party/blink/renderer/modules/payments/payments_validators.cc
when serializing data larger than than 1MB into JSON.

Reason for reland: Split out the large data serialization tests into a
separate patches for fixing the JSON serialization issue separately:
 1) https://crrev.com/c/2348411
 2) https://crrev.com/c/2348540
 3) https://crrev.com/c/2346973

Original change's description:
> [Web Payment] Secure payment confirmation Blink parameters.
>
> Before this patch, requesting "secure-payment-confirmation" method in
> PaymentRequest API would not validate the parameters.
>
> This patch adds the "secure-payment-confirmation" parameters to
> PaymentRequest API and validates them, if the
> "SecurePaymentConfirmation" feature is enabled.
>
> After this patch, PaymentRequest API validates the secure payment
> confirmation parameters.
>
> Explainer: https://github.com/rsolomakhin/secure-payment-confirmation
> Feature status: https://chromestatus.com/feature/5702310124584960
> Intent to prototype:
> https://groups.google.com/a/chromium.org/g/blink-dev/c/myUR5gyd5Js/m/iELL67NQAgAJ
>
> Bug: 1110324
> Change-Id: I7a4b6f9da1b3e1d7604bcd9e4733dc789f77ec4b
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2341163
> Reviewed-by: Nick Burris <[email protected]>
> Reviewed-by: Kouhei Ueno <[email protected]>
> Auto-Submit: Rouslan Solomakhin <[email protected]>
> Commit-Queue: Rouslan Solomakhin <[email protected]>
> Cr-Commit-Position: refs/heads/master@{#796792}

TBR=ellyjones, nburris, kouhei

Bug: 1110324, 1115091
Change-Id: I05ebdbfe234d3bee8ceb2016dafc5679a9a28ef9
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2348410
Commit-Queue: Rouslan Solomakhin <[email protected]>
Reviewed-by: Rouslan Solomakhin <[email protected]>
Reviewed-by: Nick Burris <[email protected]>
Cr-Commit-Position: refs/heads/master@{#796954}

Per spec (https://drafts.csswg.org/css-transitions-2/#event-dispatch)
a transitioncancel should be fired on the following phase change

 not idle and not after --> idle

Previously we were only test for not idle --> idle, meaning that a
finished transition could erroneously trip a transitioncancel.

Bug: 979556
Change-Id: Iecd6c8d2a7ca4564925910f092fca4c78b6d79da
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2348416
Commit-Queue: Kevin Ellis <[email protected]>
Reviewed-by: Xida Chen <[email protected]>
Cr-Commit-Position: refs/heads/master@{#796973}

Implements the spec change in
w3c/csswg-drafts#5257

This allows respecting the aspect-ratio even when we're computing the
block size from an aspect ratio with a max-height specified.

Bug: 1104360
Change-Id: Ibdff4bf4a67aa0897fd24416fed67bbb2f1ef9ea
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2308999
Commit-Queue: Christian Biesinger <[email protected]>
Reviewed-by: Ian Kilpatrick <[email protected]>
Auto-Submit: Christian Biesinger <[email protected]>
Cr-Commit-Position: refs/heads/master@{#797043}

Fixes flakiness due to a cause it is not supposed to be testing for.

Bug: 1115277
Change-Id: I41d43339913577ef20cabf528d6d1934baab3197
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2350488
Commit-Queue: George Steel <[email protected]>
Commit-Queue: Alex Russell <[email protected]>
Reviewed-by: Alex Russell <[email protected]>
Cr-Commit-Position: refs/heads/master@{#797042}

TBR: [email protected]
Bug: 1114013
Change-Id: I41b8e08e80f1d8a6ecd9e293a12a5e8f83c2bec2
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2348205
Commit-Queue: Ian Kilpatrick <[email protected]>
Reviewed-by: Ian Kilpatrick <[email protected]>
Reviewed-by: Christian Biesinger <[email protected]>
Cr-Commit-Position: refs/heads/master@{#797086}

Differential Revision: https://phabricator.services.mozilla.com/D83799

bugzilla-url: https://bugzilla.mozilla.org/show_bug.cgi?id=1652973
gecko-commit: 53cdfde742e498d81a9d1717e6a425c5b2db413c
gecko-integration-branch: autoland
gecko-reviewers: jgraham

Now that each driver supports the new window command, we can make use of it
as replacement for the create_window fixture.

Given that there is no need to check the type of the new window,
this implementation doesn't add a check for that.

Differential Revision: https://phabricator.services.mozilla.com/D83800

bugzilla-url: https://bugzilla.mozilla.org/show_bug.cgi?id=1652973
gecko-commit: 7ac192dc5d64a3e64d2a223053dba0cd6788157d
gecko-integration-branch: autoland
gecko-reviewers: jgraham

Differential Revision: https://phabricator.services.mozilla.com/D84109

bugzilla-url: https://bugzilla.mozilla.org/show_bug.cgi?id=1652973
gecko-commit: fb64d0a7a99c7835331a961de1585ca9ab308e96
gecko-integration-branch: autoland
gecko-reviewers: jgraham

This patch adds a test for JSON serialization of a very long string
being passed into PaymentRequest API.

Bug: 1110324, 1115091
Change-Id: I7966d71c3752d1e4378a54f534a1915fb8fade39
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2348411
Commit-Queue: Rouslan Solomakhin <[email protected]>
Reviewed-by: Nick Burris <[email protected]>
Cr-Commit-Position: refs/heads/master@{#797165}

Source: https://github.com/tidoust/reffy-reports/blob/c69ca2a/ed/idl/web-animations.idl
Build: https://travis-ci.org/tidoust/reffy-reports/builds/717160531

Bug: 1111134
Change-Id: I89c6b229c8871814faabadd08e3963adbe66e53e
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2348239
Reviewed-by: Kenichi Ishibashi <[email protected]>
Reviewed-by: Hiroki Nakagawa <[email protected]>
Commit-Queue: Hiroshige Hayashizaki <[email protected]>
Cr-Commit-Position: refs/heads/master@{#797197}

This adapts Chromium to the changes introduced by
https://webrtc-review.googlesource.com/c/src/+/180580, and exposes
the stop() method behind a flag.

Intent to implement and ship thread:
https://groups.google.com/a/chromium.org/g/blink-dev/c/dAeG7K7HPG8

Bug: chromium:980879
Change-Id: Iea33c0b915b5ee6ce5cab10ce98f9b906d7d9e39
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2344605
Commit-Queue: Harald Alvestrand <[email protected]>
Reviewed-by: Guido Urdaneta <[email protected]>
Reviewed-by: Kentaro Hara <[email protected]>
Cr-Commit-Position: refs/heads/master@{#797206}

This reverts commit 4fb20b8a1f2f127c6ce84ce85d1f8e4bf5299e31.

Reason for revert: Suspect for failures on WebKit Linux ASAN; https://crbug.com/1115526

Original change's description:
> [Web Payment] Very long instrumentId string test.
>
> This patch adds a test for JSON serialization of a very long string
> being passed into PaymentRequest API.
>
> Bug: 1110324, 1115091
> Change-Id: I7966d71c3752d1e4378a54f534a1915fb8fade39
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2348411
> Commit-Queue: Rouslan Solomakhin <[email protected]>
> Reviewed-by: Nick Burris <[email protected]>
> Cr-Commit-Position: refs/heads/master@{#797165}

[email protected],[email protected]

Change-Id: Ieef62c51b197a87a04b4bfc4fd6c91f8bcea85a6
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: 1110324
Bug: 1115091
Bug: 1115526
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2352549
Reviewed-by: Stephen McGruer <[email protected]>
Commit-Queue: Stephen McGruer <[email protected]>
Cr-Commit-Position: refs/heads/master@{#797244}

Version as reported by safaridriver --version:
Included with Safari Technology Preview (Release 111, 15610.1.23.1.2)

Source: https://developer.apple.com/safari/download/
Build: https://dev.azure.com/foolip/safari-technology-preview-updater/_build/results?buildId=2069&view=logs

Co-authored-by: Sam Sneddon <[email protected]>

…ative.idl

This CL adds everything in TrustedTypePolicyFactory that changes were made to
the spec that weren't reflected in the IDL tests description in
`trusted-types.tentative.idl` test file.

Change-Id: Iac7c68439a449d5437783daa7640ffcb1d3694a6
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2290090
Commit-Queue: Yifan Luo <[email protected]>
Reviewed-by: Daniel Vogelheim <[email protected]>
Reviewed-by: Mike West <[email protected]>
Cr-Commit-Position: refs/heads/master@{#797337}

This reverts commit 008d0a49de0c0fa5fb0a566827e6d964697516d3.

Reason for revert: Decision on w3c/csswg-drafts#5400

Original change's description:
> Fix Animation::playState
>
> w3c/csswg-drafts#5400
>
> There is the issue that led to the spec change for play state:
>
> w3c/csswg-drafts#2066
>
> The issue addresses undesirable toggling of the play state due to
> inactive timelines, but does not seem to adequately address having an
> unresolved (null) timeline.
>
>
> Bug: 1113382
> Change-Id: I33032ee1e9de10d3dce4958fce599bd98391328f
> Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2339676
> Reviewed-by: Anders Hartvoll Ruud <[email protected]>
> Commit-Queue: Kevin Ellis <[email protected]>
> Cr-Commit-Position: refs/heads/master@{#796403}

[email protected],[email protected]

# Not skipping CQ checks because original CL landed > 1 day ago.

Bug: 1113382
Change-Id: I24aa0ef5c62f839573a70f063c1f25ce4281b2e5
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2352513
Reviewed-by: Kevin Ellis <[email protected]>
Commit-Queue: Kevin Ellis <[email protected]>
Cr-Commit-Position: refs/heads/master@{#797351}

The CL modifies URLLoaderFactoryParamsHelper::CreateForRendererProcess
so that it sets |request_initiator_site_lock| of the process-wide
factory to an opaque origin.  This change is based on
https://crbug.com/1105794 which shows that there are indeed no cases in
the wild where a process-wide factory is used with a non-opaque
initiator origin.  This intuitively makes sense - the process-wide
factory should only be used in frames that have not yet committed any
navigation and such frame should not (yet) have a non-opaque origin.

Additionally, thanks to the change described above, there is no longer a
need for site/eTLD+1 comparisons in VerifyRequestInitiatorLock nor for
SiteInstanceImpl::GetRequestInitiatorSiteLock.

Fixed: 1098938
Change-Id: I9e8b53139a2418636c84e783f86ea8d7be34eed9
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2274591
Reviewed-by: Kinuko Yasuda <[email protected]>
Commit-Queue: Łukasz Anforowicz <[email protected]>
Cr-Commit-Position: refs/heads/master@{#797542}

Fixes #24924

Follows whatwg/html#5809.

Bug: 1112857
Change-Id: I2c76e95ecb139d3359877e71c6f3713f4d9f3c37
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2348694
Reviewed-by: Philip Rogers <[email protected]>
Commit-Queue: Annie Sullivan <[email protected]>
Cr-Commit-Position: refs/heads/master@{#797698}

See [1] for context. This CL adds the shadowRoot attribute to ElementInternals,
which allows custom elements to access their own shadow roots, including in
the case that the shadowRoot is closed. Without this capability, custom
elements that use closed declarative shadow roots would have no ability to
retrieve the contents of their own shadow root.

Bug: 1042130

[1] WICG/webcomponents#871 (comment)

Change-Id: I33606144bec3c27600d2e67f108d0344f7696dd2
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2350739
Commit-Queue: Mason Freed <[email protected]>
Auto-Submit: Mason Freed <[email protected]>
Reviewed-by: Kouhei Ueno <[email protected]>
Cr-Commit-Position: refs/heads/master@{#797729}

…n3 (#24685)

See whatwg/html#5805; this tests the scenario there plus others.

Currently the inline rAF loop gets restarted before the page is
notified that the immersive session has ended. This causes some flaky
test cases where the inline callback is serviced before the page knows
that it's okay, since the immersive session has ended. This change
splits the FrameProvider's OnSessionEnded logic such that it can be
notified of the fact that there is no longer an immersive session before
the page is (so that the page is not blocked from requesting a new
immersive session), while only re-starting the inline rAF loops after
the page has been notified.

Fixed: 1062195
Change-Id: I2ccf63d476b7f857bc53e9c15bedc07a13fdfed6
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2354537
Reviewed-by: Brandon Jones <[email protected]>
Commit-Queue: Alexander Cooper <[email protected]>
Cr-Commit-Position: refs/heads/master@{#797812}

Prior to this change, effects are always given the current phase of the
timeline associated with them. This fails in certain situations such as
pausing and then scrolling, setting current time with an inactive
timeline, and many other scenarios where hold_time is set.

In order to fix this, we have introduced hold_phase. Anytime hold_time
is updated, hold_phase will also be updated with the appropriate phase.
Sometimes the phase is set explicitly and other times we pull from the
timeline current time.

Animation effects will be given the animation hold_phase if it is
populated, otherwise it will calculate its current phase (similar to how
current time is calculated).

The WebAnimations spec is being updated to reflect this change.

Bug: 1046833
Change-Id: I4bf1e42eaab684c18829a79acc1ab8911ec893af
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2135336
Commit-Queue: Jordan Taylor <[email protected]>
Reviewed-by: Kevin Ellis <[email protected]>
Reviewed-by: Majid Valipour <[email protected]>
Cr-Commit-Position: refs/heads/master@{#797835}

We now have established a formal way to do this (more documentation will
be available in Chromium, as this is Chrome-only) and should no longer
accept these auto-generated files in WPT.

Related to #16455.

Implements putAllValues where any errors cause the whole request
to fail, instead of just the value/s that failed.

Explainer: https://github.com/nums11/idb-putall/blob/master/Explainer.md

Design Doc: https://docs.google.com/document/d/1wawQ8Pl-Vi6GQN5-y2UVkcghipcOGg0WRAC0ZyBkezg/edit

I2P: https://groups.google.com/a/chromium.org/g/blink-dev/c/TPd_rtgO3_k/m/qZmPy1dfAgAJ

Change-Id: I976db2fe4c441f0ae288e63deb4c69c148721a18
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2324146
Reviewed-by: Christian Biesinger <[email protected]>
Reviewed-by: Joshua Bell <[email protected]>
Reviewed-by: Chris Palmer <[email protected]>
Reviewed-by: Daniel Murphy <[email protected]>
Reviewed-by: Olivier Yiptong <[email protected]>
Commit-Queue: Daniel Murphy <[email protected]>
Cr-Commit-Position: refs/heads/master@{#797837}

Updates the README in WPT to reflect this, and Chromium's roll script so
that it doesn't roll into WPT.

And, unrelatedly, removes JusSn from the suggested_reviewers as he is
no longer actively working on WebGPU.

Rationale:

These tests test an unstable, constantly changing API. Because of this,
it's not practical for browser implementations to try to sync on a
single version of the test suite which is auto-imported/exported from
WPT. Auto-imports in any browser will necessarily result in test
failures just because the tests have changed to match some API change.

Instead, for the moment, browsers should pin ("vendor") a particular
version of the WebGPU CTS, and may include a built version of it in
their equivalent of "wpt_internal" for automated testing.

Discussion: #24699

Bug: None
Change-Id: Iaabd81ef646404ab7f6e8cc19a884b15f5306bca
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2353550
Reviewed-by: Rakib Hasan <[email protected]>
Reviewed-by: Austin Eng <[email protected]>
Commit-Queue: Kai Ninomiya <[email protected]>
Cr-Commit-Position: refs/heads/master@{#797946}

This CL adds basic tests, verifying VideoTrackReader construction, start
and stop.

Bug: 1094194
Change-Id: I945b49c83b30fb18e1b92f2263fbdb050a431d02
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2353533
Reviewed-by: Chrome Cunningham <[email protected]>
Commit-Queue: Thomas Guilbert <[email protected]>
Cr-Commit-Position: refs/heads/master@{#797952}

Updating timelines can enqueue tasks related to ready and finish
promises which should run before animation events and updated
rendering.

Bug: 1116020
Change-Id: Ica1927bb3921eef574abeaca23bb198dda658ccd
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2354164
Reviewed-by: Kevin Ellis <[email protected]>
Commit-Queue: Robert Flack <[email protected]>
Cr-Commit-Position: refs/heads/master@{#798259}

Background: transformed rasterization is used to render sharp text in a
composited layer with fractional transform. It snaps the composited
transform of the layer to whole pixels, then applies the fraction
during rasterization.

Previously we disallowed transformed rasterization in blink for directly
composited layers (excluding backface-visibility:hidden and trivial 3d
transforms such as translateZ(0)), especially for will-change:transform
because change of fractional transform would cause re-rasterization
which would defeat the purpose of will-change:transform.

Now for a layer that was not eligible for transformed rasterization,
calculate the initial raster translation which achieves sharp text
rendering with the initial transform. When the transform of the layer
changes, keep its original raster translation to avoid invalidation
of tilings, to ensure animation performance. The rendered result will
be blurry during animation but the initial rendering result is sharp.

Remove blink-side transformed rasterization conditions and let cc
decide it based on screen and draw transforms of the layers.

Bug: 1111195
Change-Id: If0455be2f29e69cd7152c951626e6e8cc143b5bd
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2351164
Reviewed-by: Philip Rogers <[email protected]>
Reviewed-by: vmpstr <[email protected]>
Commit-Queue: Xianzhu Wang <[email protected]>
Cr-Commit-Position: refs/heads/master@{#798268}

For same-process iframes, propagate the Document-level IsXrOverlay flag,
and use the correct element as the overlay layer.

This moves the document->SetIsXrOverlay setters from xr_system to
fullscreen to ensure it gets applied consistently. There are multiple
documents that need updating when iframes are being used. (Handling
out-of-process iframes needs additional logic that's being added
in a followup CL.)

Bug: 1101193
Change-Id: Icf80e65e1b927888c6140acbc998ed90d469b5e1
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2283823
Commit-Queue: Klaus Weidner <[email protected]>
Reviewed-by: Philip Jägenstedt <[email protected]>
Reviewed-by: Alexander Cooper <[email protected]>
Cr-Commit-Position: refs/heads/master@{#798312}

Propagate the "is_xr_overlay" fullscreen request type flag for OOPIF
requests via mojo, and ensure that visibility is set correctly for
out of process container iframe elements.

Note for testing: Currently, Chrome on Android only enforces site
isolation for sites with login active, and not on low memory devices.
Run with the --site-per-process command line flag to override this.

Bug: 1101193
Change-Id: Ic8736e90f28fbda2d6cb1f665cbd0804c3f38cc3
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2283968
Commit-Queue: Klaus Weidner <[email protected]>
Reviewed-by: Philip Jägenstedt <[email protected]>
Reviewed-by: Philip Rogers <[email protected]>
Reviewed-by: Daniel Cheng <[email protected]>
Cr-Commit-Position: refs/heads/master@{#798382}

The frame selection changes can run script, so we need to check for the
frame itself getting destroyed. This fixes the crash as reported.

Additionally, the document check it does for pointer capture is
incorrect, it should use the composed, not uncomposed doc, so that it
works in shadow dom.

Differential Revision: https://phabricator.services.mozilla.com/D87097

bugzilla-url: https://bugzilla.mozilla.org/show_bug.cgi?id=1386418
gecko-commit: e140c6cd2eb944579ec99458638879c9b8501920
gecko-integration-branch: autoland
gecko-reviewers: dholbert

Implement the mpadded element [1]. This CL also renames LayoutNGMathMLSquareRoot to LayoutNGMathMLBlockWithAnonymousMrow, since now it is not just used by msqrt.

[1] https://mathml-refresh.github.io/mathml-core/#adjust-space-around-content-mpadded

Bug: 6606
Change-Id: I8243bb8e5fb0975a407bf12275487434283cfe4c
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2288388
Commit-Queue: Rob Buis <[email protected]>
Reviewed-by: Ian Kilpatrick <[email protected]>
Cr-Commit-Position: refs/heads/master@{#798451}

This test worked on Mozilla due to the adoption of orient="vertical"
and the native sizing of the range control. It failed on Chromium
because min-content was less than 30px. This new version appears
to pass on both Chromium and Mozilla.

Fixed: 1108796
Change-Id: Ie5cbaaff6c78f7658931e8d63860383e3501f5ad
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2355567
Auto-Submit: Mason Freed <[email protected]>
Commit-Queue: Ian Kilpatrick <[email protected]>
Reviewed-by: Ian Kilpatrick <[email protected]>
Cr-Commit-Position: refs/heads/master@{#798460}

- don't clear the error callback prior to calling it
- implement encodeQueueSize to signal backpressure
- add extra error logging for VpxVideoEncoder
- disable vp9 profiles 1 and 3 (not yet implemented)

And opportunistically mark a number of inputs
in IDL as required
- timestamp (its required for creation of VideoFrame internally)
- error and output callbacks (duh).

Remove ? from duration input dictionary. Its not required,
no need to null.

Bug: 1094166, 1116617

Change-Id: I9577bc42d1f24c2ec14885401e87337ba0937e43
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2355276
Commit-Queue: Chrome Cunningham <[email protected]>
Reviewed-by: Thomas Guilbert <[email protected]>
Cr-Commit-Position: refs/heads/master@{#798473}

The only relative url works fine regardless of the base anyway.

Differential Revision: https://phabricator.services.mozilla.com/D87179

bugzilla-url: https://bugzilla.mozilla.org/show_bug.cgi?id=1659249
gecko-commit: ffc01c0f13a841719e86d032ac27fe9dbc1fc5f7
gecko-integration-branch: autoland
gecko-reviewers: mats

…25008)

Most of the changes made by:

> git grep -lF os.path.split | xargs sed -i '' 's/os\.path\.split(\(.*\))\[0\]/os.path.dirname(\1)/g'

That gets some things wrong, so all changes were vetted and some fixed
or simplified.

…ords

Per the resolution here:
w3c/csswg-drafts#2943 (comment)

Differential Revision: https://phabricator.services.mozilla.com/D86848

bugzilla-url: https://bugzilla.mozilla.org/show_bug.cgi?id=1658780
gecko-commit: f700072e4306e5c87fcc84752fcad59fe8f13d52
gecko-integration-branch: autoland
gecko-reviewers: emilio

Source: https://github.com/tidoust/reffy-reports/blob/05fb168/ed/idl/uievents.idl
Build: https://travis-ci.org/tidoust/reffy-reports/builds/718526289

See #23918

Follows whatwg/html#5302.

Corresponds to whatwg/html#5827.

There were harness errors in Edge, Firefox and Safari:
https://wpt.fyi/results/html/canvas/element/manual/wide-gamut-canvas/canvas-createImageBitmap-e_srgb.html?run_id=657990001&run_id=659880001&run_id=633480001&run_id=663850001

The cause for that was the convertToBlob call not correctly wrapped in a
step in testCreateImageBitmapFromColorManagedBlob.

That was only the first of multiple errors fixed here:
 - not handling image or video decode failures, causing timeouts
 - failures at the end of testCreateImageBitmapFromColorManagedBlob
   would not have failed the test because promise rejections were
   not handled
 - failing in cryptic ways if ImageData dataUnion isn't supported

These issues were fixed while also modernizing the tests to use more
modern APIs where available, like img.decode() and fetch().

Change-Id: Ia8d0086a2d216dd291c9429460f8a79f8f4fd1c6
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2359991
Reviewed-by: Wan-Teh Chang <[email protected]>
Reviewed-by: Fernando Serboncini <[email protected]>
Commit-Queue: Philip Jägenstedt <[email protected]>
Cr-Commit-Position: refs/heads/master@{#798674}

This patch extracts out the stacking context grouping properties from
ComputedStyle::HasGroupingProperty, and uses this for the stacking
context decision. This has the effect of not creating a stacking context
for clip which, unlike clip-path, does not create a stacking context:
  https://drafts.fxtf.org/css-masking-1/#propdef-clip

Bug: 1115003
Change-Id: I4bd5d3c61a2bb34a56a089cbeaca260fc9bf2eb4
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2358473
Commit-Queue: Philip Rogers <[email protected]>
Commit-Queue: Xianzhu Wang <[email protected]>
Auto-Submit: Philip Rogers <[email protected]>
Reviewed-by: Xianzhu Wang <[email protected]>
Cr-Commit-Position: refs/heads/master@{#798712}

And remove unused mojom.js files.

Bug: 1094512
Change-Id: Ibe8503d72fff30b96248c14647b65481c06c7b2f
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2350482
Commit-Queue: Robert Ma <[email protected]>
Reviewed-by: Reilly Grant <[email protected]>
Auto-Submit: Robert Ma <[email protected]>
Cr-Commit-Position: refs/heads/master@{#798691}

Source: https://github.com/tidoust/reffy-reports/blob/3b17e3d/ed/idl/WebIDL.idl
Build: https://travis-ci.org/tidoust/reffy-reports/builds/718725371

This reports data: and sandboxed iframes as always origin-isolated,
according to the window.originIsolationRestricted property. This is
done regardless of the process isolation state or the presence/absence
of the Origin-Isolation header.

The web platform tests were revamped in the following ways:

* Sandboxed iframes and data: URL tests were split into -yes and -no
  variants, to ensure that the parent frame didn't influence the
  result. For sandboxed iframes in particular, the -no variant was not
  tested previously, and was failing until after the code changes here.

* Additional tests for javascript: URLs were introduced, as an earlier
  version of this CL would trigger a DCHECK on opaque-origin pages
  navigating to javascript: URLs.

* Shared code was factored out. The .html files are now just wrappers
  that delegate to resources/*-test.mjs files.

Bug: 1095653
Change-Id: I69a1f388250d1364dd33c365c936a76546add5b2
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2334103
Reviewed-by: Nate Chapin <[email protected]>
Reviewed-by: Alex Moshchuk <[email protected]>
Reviewed-by: Arthur Sonzogni <[email protected]>
Commit-Queue: Domenic Denicola <[email protected]>
Cr-Commit-Position: refs/heads/master@{#798741}

Before this patch, invoking PaymentRequest API with secure payment
confirmation method identifier and options to request shipping or
contact information was considered valid.

This patch throws RangeError in PaymentRequest constructor if secure
payment confirmation method identifier is requested together with
shipping or contact information.

After this patch, it's not possible to request secure payment
confirmation method together with shipping or contact information.

Bug: 2348410
Change-Id: I3e01f9b758645a1653533f41da9688c453199b03
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2359554
Reviewed-by: Liquan (Max) Gu <[email protected]>
Commit-Queue: Rouslan Solomakhin <[email protected]>
Cr-Commit-Position: refs/heads/master@{#798755}

Change the test to follow the proposed change in w3c/FileAPI#41,
now that all browsers agree to not replace /.

Depends on D86981

Differential Revision: https://phabricator.services.mozilla.com/D87112

bugzilla-url: https://bugzilla.mozilla.org/show_bug.cgi?id=1650607
gecko-commit: 365ae38a5bca68e4d8c8cf25b8f5230e306b0f09
gecko-integration-branch: autoland
gecko-reviewers: baku

For elements with will-change:transform or directly composited images
the raster scale is not updated in response to transform scale
changes. Once the scale gets especially large, the interest rect
computed by blink starts to become less than the layer bounds.
This ends up setting a recorded_viewport_ on RasterSource that is also
less than the layer bounds. If the interest rect gets small enough, it's
possible that tiles that are in view are not completely covered by the
RasterSource. This means that we won't create tiles (much less assign
them GPU memory and schedule them) in PictureLayerTiling::CreateTile. We
did not see this with the previous implementation of directly composited
images, since the content layer never had an interest rect computed.

This change fixes this specifically for directly composited images by
always considering its raster source as covering the entire layer. This
is the definition of directly composited images (has a DrawImageRectOp
that draws into the entire layer). This allows the in view tiles to be
created, then rasterized and activated and the image no longer
disappears.

I looked into fixing the general case (see comments in the bug) but
given the regression for the images scenario and the fact that we'd like
to get a fix merged for 85, I took an approach to just fix directly
composited images.

Bug: 1110141
Change-Id: Ice3cf25508585f116edd7317718292ec9ad8a7ef
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/2358888
Reviewed-by: Philip Rogers <[email protected]>
Reviewed-by: vmpstr <[email protected]>
Commit-Queue: Daniel Libby <[email protected]>
Cr-Commit-Position: refs/heads/master@{#798836}